Organizational Behavior and Human Decision Processes 152 (2019) 1–10

Contents lists available at ScienceDirect

Organizational Behavior and Human Decision Processes

journal homepage: www.elsevier.com/locate/obhdp

Leader power, power stability, and interpersonal trust T

a,⁎ b b c
Marlon Mooijman , Wilco W. van Dijk , Eric van Dijk , Naomi Ellemers
a
Northwestern University, United States
b
Leiden University, the Netherlands
c
Utrecht University, the Netherlands

A R T I C LE I N FO A B S T R A C T

Keywords: We examine the conditions under which power decreases trust and the process by which this occurs. Three
Power experiments and a field study revealed that occupying an unstable power position decreases trust as it raises
Distrust power holders’ concerns about losing power. We replicate this finding across studies differing in measures and
Power-loss concerns design, using different samples, and using different measurements and manipulations of power, power stability,
and trust. We discuss the organizational and theoretical implications of this work for current insights on power
and trust.

1. Introduction leaders’ trust, productivity, reciprocity, and cooperation break down

Trusting can be essential for leaders. Teams perform better (e.g.,
team members engage in more OCB behaviors), organizations experi-
ence fewer struggles (e.g., employees are less likely to quit), and gov-
ernments collect more taxes when people feel trusted by their leaders
(Brower, Lester, Korsgaard, & Dineen, 2009; Feld & Frey, 2002; Lau &
Lam, 2008). Yet, leaders do not always trust. Recent research suggests
that due to their power, leaders at times distrust those they interact
with, and engage in behaviors that signal this distrust to these inter-
action partners (e.g., by trying to control them; Inesi, Gruenfeld, &
Galinsky, 2012; Mooijman, Van Dijk, Ellemers, & Van Dijk, 2015;
Mooijman, Van Dijk, Van Dijk, & Ellemers, 2017; Mooijman & Graham,
2018; Schilke, Reimann, & Cook, 2015). But when and why does lea-
ders’ power decrease their trust in their interaction partners? And how
can this be prevented? In the present research, we examine the condi-
tions under which power decreases trust and the process by which this
occurs. We propose that occupying an unstable power position de-
creases leaders’ trust through raising power-loss concerns. We further
propose that alleviating these concerns attenuates the extent to which
power decreases trust.
Understanding how having power affects trusting attitudes and
behaviors is important, as trust is vital in hierarchical contexts. For
instance, managers must trust employees’ willingness to comply with
their instructions and do what is best for the company (Berger,
Rosenholtz, & Zelditch, 1980) without resorting to punitive systems for
enforcing cooperation (Kirchler, Kogler, & Muelbacher, 2014). Without
(Brown, Gray, McHardy, & Taylor, 2015). Leaders who develop trusting
relationships with employees (and others in the organization) are more
effective but these leaders are often more powerful than others (De
Jong, Van der Vegt, & Molleman, 2007). Explaining when and why
power undermines trust is thus important because it helps understand
when and why leaders’ lack of trust may proliferate in hierarchical
contexts and how this can be prevented.
2. Trust
Trust can be defined as: “the willingness to be vulnerable to the
actions of another party based on the expectation that the other will
perform a particular action important to the trustor, irrespective of the
ability to monitor or control that other party” (Mayer, Davis, &
Schoorman, 1995, p. 712). Central to this definition of trust is that the
willingness to be vulnerable to others is based on the expectation that
these others have benevolent intentions (see for similar definitions,
Lewicki, Tomlinson, & Gillespie, 2006; Zand, 1997). Trust thus has two
components: (a) an expectation of goodwill (benevolence) from others,
and (b) the willingness to open oneself up for potential exploitation
based on this expectation of goodwill. For instance, expecting others to
further one’s interest (i.e., goodwill) increases the likelihood that in-
dividuals make themselves vulnerable to others (i.e., trust; Lount &
Petitt, 2012). Trust can be distinguished from assurance, which is a
willingness to be vulnerable to others that is not based on an expecta-
tion of benevolence (e.g., when one believes others will cooperate in

⁎
Corresponding author at: Management and Organizations Department, Kellogg School of Management, Northwestern University, 2211 Campus Dr, Evanston, IL
60208, United States.
E-mail address: marlon.mooijman@kellogg.northwestern.edu (M. Mooijman).

https://doi.org/10.1016/j.obhdp.2019.03.009
Received 1 February 2016; Received in revised form 6 March 2019; Accepted 20 March 2019
0749-5978/ © 2019 Elsevier Inc. All rights reserved.
M. Mooijman, et al.
fear of punishment; Yamagishi & Yamagishi, 1994). In contrast to as-
surance, trust entails the expectation that others will act in one’s in-
terest, even though this is not in their own self-interest (e.g., trust in
social-dilemma situations; Mulder, Van Dijk, De Cremer, & Wilke,
2006). We similarly operationalize trust as an expectation of goodwill
resulting in an increased willingness to be vulnerable to others in si-
tuations with conflicting interests: for instance, managers have to trust
subordinates to use organizational resources to pursue workplace goals
and be productive instead of use these resources for selfish goals (e.g.,
take office supplies home).
Previous research has demonstrated that trust can be affected by
one’s position in the power hierarchy. Using U.S. president Johnson’s
office conversations, telephone calls, and interviews with personal
friends, Kramer and Gavrieli (2004) suggested that president Johnson
was suspicious of others—and suggested that this was in part due to his
position of power. Mooijman et al. (2015) and Schilke et al. (2015)
demonstrated experimentally that being assigned a position of power
(versus powerlessness) in itself can be sufficient to decrease an in-
dividual’s trust in others. Although these studies are informative (for
effects of status on trust, see Lount & Petitt, 2012), they do not delineate
the specific conditions under which power does and does not under-
mine trust, nor do they elucidate whether or how this might be pre-
vented. Thus, it is unclear when power decreases trust and why this
occurs. The current research extends prior findings and addresses this
issue by investigating the role of power stability and power-loss con-
cerns.
3. Power-loss concerns
Power can be defined as having asymmetric control over valuable
resources (Anderson & Brion, 2014; Magee & Galinsky, 2008). These
resources can be monetary (e.g., ability to grant salaries, bonuses),
social (e.g., ability to grant inclusion in respected teams), or physical
(e.g., ability to grant office spaces). Power is inherently a relational
construct, as one party (leader) controls what the other party in the
relationship (subordinate) desires. Being the more powerful party in a
relationship is beneficial. For instance, power holders can disregard
others’ desires and feelings (Goodwin, Gubin, Fiske, & Yzerbyt, 2000;
Van Kleef et al., 2008, and focus on pursuing their own instead of
others’ goals (Galinsky, Gruenfeld, & Magee, 2003; Guinote, 2007;
Lammers, Stoker, Jordan, Pollmann, & Stapel, 2011). Power boosts
individuals’ self-esteem (Wojciszke & Kujalowicz-Struzynska, 2007) and
leads them to express more positive—approach related—emotions (e.g.,
amusement and happiness) and sometimes less negative—inhibition
related—emotions (e.g., embarrassment and anger; Bombari, Schmid, &
Bachmann, 2017; Keltner, Anderson, & Gruenfeld, 2003).
Given the benefits of power, individuals in organizations often do
not like to lose power once it is gained. Research suggests that when
individuals obtain a position that enables them to make group decisions
(i.e., power), the majority chooses to maintain their power position,
even when delegating this position to others would be in their economic
self-interest (Fehr, Herz, & Wilkening, 2013). The prospect of losing
power also raises individuals’ physiological responses indicating the
experience of threat (e.g., as expressed through heart rate and blood
pressure; Scheepers, Röell, & Ellemers, 2015) and when powerful lea-
ders perceive their power to be threatened they are more likely to
create divisions among less powerful individuals in such a way that it
prevents them from forming alliances (e.g., divide and conquer; Case &
Maner, 2014). Moreover, the finding that powerful individuals pay
more attention (and seek proximity) to individuals whom they perceive
as a threat to their power (Maner & Mead, 2010) suggests that the
potential of losing power is perceived as threatening and aversive.
4. Power-loss concerns and trust
One way in which leaders can lose their power is by trusting
2
Organizational Behavior and Human Decision Processes 152 (2019) 1–10
individuals who can take power away from them or gain power relative
to them. Opening oneself up for potential exploitation can be expressed
in resource sharing (i.e., power sharing within a relational context) that
entails a risk of losing resource control (Kramer, 1999; Mayer et al.,
1995; Zand, 1997). The Trust Game nicely illustrates this in an eco-
nomic exchange context. In the Trust Game, trusting others means
giving away control over a monetary resource to others, while ex-
pecting these others to give this control back at the end of the game.
When one’s trust is vindicated, one keeps control over the money; when
it is violated, one loses control over the money (Berg, Dickaut, &
McCabe, 1995). Within the context of power in a relationship, trust in
the Trust Game means giving away your power to the other party (as
the other party now has more control over resources than before). Ex-
pecting not-so-benevolent intentions from others may lead people to
protect their control over resources to a greater degree—as expecting
others to be untrustworthy makes one less willing to give up resource-
control and thus less likely to potentially undermine one’s power
(Mayer et al., 1995). This reasoning suggests that those who are more
concerned about losing their power are less inclined to place trust in
individuals who potentially compete with them for resources, as this
would make them vulnerable to power loss.
Importantly, power is not always stable in real life, with salient
opportunities for power loss (Kim, Shin, & Lee, 2015). The stability of a
power hierarchy is an important determinant of potential power loss
(Jordan, Sivanathan, & Galinsky, 2011): leaders such as managers can
experience feelings of instability with regard to their own position (e.g.,
being demoted to a less powerful role) and with regard to the organi-
zational resources they control (e.g., shrinking organizational budgets).
Leaders can also experience stability, such that leaders feel secure in
their position (e.g., unlikely to be fired or demoted) and the organiza-
tional resources that they control (e.g., secure organizational budgets).
As such, leaders are likely to be concerned about power loss, and trust
less, when they perceive the power hierarchy as unstable. Unstable
power hierarchies may decrease the extent to which power holders are
willing to trust individuals who can potentially take resources away
from them and, in doing so, improve their power position relative to the
leader. For instance, within organizations, power is often manifested in
control of vital resources such as organizational budgets. Sharing those
resources with others within the organization can sometimes lead to
relative power loss (e.g., when these others become less dependent on
the leader). If having an unstable power position evokes power-loss
concerns and not trusting individuals who can threaten one’s power is
(perceived as) a way to avoid losing power to these individuals, then
unstable power should decrease trust through raising these power-loss
concerns.
These predictions connect the trust literature with the literature on
how unstable power can increase physiological threat responses
(Scheepers et al., 2015) and moves theorizing beyond a simple main
effect of power on trust (Mooijman et al., 2015; Schilke et al., 2015).
Specifically, Scheepers et al. (2015) demonstrated the impact of un-
stable power on cardiovascular response patterns (e.g., heart rate, blood
pressure), consistent with the idea that unstable power triggers a con-
cern for power loss and a subsequent decrease in trust. Our theorizing
also suggests that unstable power has different effects on trust than on
other forms of non-social risk-taking (e.g., unstable power can make
people more risk-taking in economic contexts; Jordan et al., 2011).
Specifically, trust indicates the willingness to take social risks and can
be distinguished from non-social risk-taking in for instance a financial
decision-making context (Houser, Schunk, & Winter, 2010). Trust is
based on your interaction partner’s potential decisions and intentions,
whereas economic risk-taking typically relies on the likelihood of out-
comes that do not involve another person (e.g., investing money with a
certain probability of return on this investment). Indeed, a person’s
likelihood of trusting others is not always (tightly) connected to a
person's economic risk attitudes (see Fetchenhauer & Dunning, 2012;
Houser et al., 2010) and the outcomes of (unwarranted) risks might be
M. Mooijman, et al.
different in social versus non-social contexts: those who take social risks
are more likely to jeopardize their standing and relations with others
(e.g., lose power to these others) compared to non-social economic risk-
taking where people might be relatively free to take risks again (e.g.,
after a bankruptcy). The current theorizing is thus primarily relevant to
trust but not necessarily economic risk-taking, as other people are less
likely to take away power holders’ resources (and thus power) when
engaging in economic risk-taking.
5. Overview of current research
We tested our central prediction in three experiments and a field
study. These studies were conducted on Amazon’s Mechanical Turk
(Studies 1 and 3), in the laboratory (Study 2), or in the field using
samples of supervisors and subordinates (Study 4). Unless indicated
otherwise, all measured variables were assessed on seven-point scales
on which participants could indicate their level of agreement
(1 = disagree completely, 7 = agree completely). All participants provided
informed consent and were debriefed, compensated, and thanked for
their participation. In each study we indicate any data exclusions, all
manipulations, and all measures. Studies 1 and 3 also included an at-
tention check but results did not differ significantly regardless of
whether participants who failed the attention check were included or
excluded from the analysis. We, therefore, report the analyses that were
conducted on the full sample, including participants who failed the
attention check1. No other studies were conducted on the impact of
unstable power on trust that are not reported in the current manu-
script2.
5.1. Study 1
In Study 1, we manipulated both the extent to which participants
controlled a large (high power) compared to a small proportion (low
power) of resources—consistent with the conceptualization of power as
asymmetric control over valuable resources (Magee & Galinsky,
2008)—and the stability of this resource control, and measured trust.
As such, we directly tested our main prediction that occupying an un-
stable power position decreases an individual’s trust in others. In ad-
dition, unstable power and the illegitimacy of power can have similar
behavioral effects (Lammers, Galinsky, Gordijn, & Otten, 2008; Maner,
Gailliot, Butz, & Peruche, 2007), with power illegitimacy defined as the
perception that one’s power is not fairly earned (Lenski, 1966; Mills,
1956). We therefore also measured perceived legitimacy of one’s power
to rule out the possibility that an observed effect of unstable power on
trust could be explained by a decrease in its perceived legitimacy (e.g.,
Lammers et al., 2008).
5.2. Method
5.2.1. Participants and design
A total of 206 Mechanical Turk participants (124 males;
Mage = 35.28 years, SDage = 11.38; see Buhrmester, Kwang, & Gosling,
2011; Buhrmester, Talaifar, & Gosling, 2018; Hauser & Schwarz, 2016,
for a discussion of Mechanical Turk as a research tool) participated for
$0.50 and were randomly assigned to a 2 (high power vs. low
power) × (stable power vs. unstable power) between-participants de-
sign. We aimed to recruit 50 participants per condition, consistent with
previous recommendations (Simmons, Nelson, & Simonsohn, 2011,
2013). A power analysis using G*Power 7 shows that this sample size
1
A total of eight (Study 1) participants failed the attention check and were
excluded from the analyses. However, the pattern of results was the same, and
still significant, when these participants were included in the analyses.
2
For data and syntax for all four studies, see the first author’s OSF page at
https://osf.io/4psv7/?view_only=8c57d9eb22b84099aa307d7e14bbaf43.
3
Organizational Behavior and Human Decision Processes 152 (2019) 1–10
provided 80% power to detect a medium sized two-way interaction
effect (f = 0.25).
5.3. Procedure
5.3.1. Power manipulation
The power manipulation was based on prior power research (Blader
& Chen, 2012). Participants were asked to place themselves in the role
of a vice-president of sales at a medium-sized firm. In the high power
[low power] condition participants read: “You are a manager who is
recognized as one of the most [least] powerful individuals within the com-
pany, and your division is widely recognized as one of the most [least]
important divisions company-wide. As a result, your division is allocated one
of the largest [smallest] budgets in the firm, and you have control over an
unusually large [small] amount of resources, compared with your colleagues
in other divisions”. Participants in both the high-power and low-power
condition thus had control over resources and could be concerned about
losing those resources.
5.3.2. Power stability manipulation
For participants in the unstable [stable] power condition, the power
manipulation also included a sentence that stated: “To stimulate pro-
ductivity across the entire firm, managers regularly change divisions within
the firm so that other managers can take their place [to stimulate pro-
ductivity across the entire firm, managers almost never change divisions
within the firm]”. This manipulation is consistent with previous ma-
nipulations of power position stability (Case & Maner, 2014; Jordan
et al., 2011).
5.3.3. Trust
Participants then indicated the extent to which they trusted other
managers of the firm: “I think that I can fully trust other managers at
the firm”, “Other managers at the firm can be trusted”, and “Other
managers at the firm always take my interests into account”,
Cronbach’s α = 0.75).
5.3.4. Manipulation checks
Perceived stability of one’s power position was measured with one
item on a seven-point scale: “My position at the firm is unstable”. To
check whether the effect of the power stability manipulation was in-
dependent of participants' overall sense of power (and to exclude this as
an alternative explanation for observed effects), we also measured
perceived power with a three-item scale: “As a manager I feel pow-
erful”, “As a manager, I feel influential”, and “As a manager, I feel in
power” (α = 0.97). Lastly, we measured perceived legitimacy: “I feel
like my position as a manager is legitimate”.
5.4. Results
5.4.1. Manipulation checks
A univariate analysis of variance (ANOVA) with power and power
stability as independent variables and perceived power as dependent
variable yielded a main effect of power, F(1, 202) = 251.23, p < .001,
η2p = .55. Participants in the high-power condition (M = 6.16,
SD = 0.87) felt more powerful than participants in the low-power
condition (M = 3.03, SD = 1.76). We observed no main effect of power
stability (F[1, 202] = 0.49, p = 0.83, η2p = .00) and no interaction ef-
fect between power and power stability, F(3, 202) = 0.21, p = .65,
η2p = .00.
In addition, an ANOVA with power and power stability as in-
dependent variables and perceived stability as dependent variable
yielded a main effect of power stability, F(1, 202) = 6.60, p = .011,
η2p = .03. Participants in the unstable-power condition (M = 4.79,
SD = 1.67) felt more unstable in their power than participants in the
stable-power condition (M = 4.31, SD = 1.99). We observed no main
effect of power (F[1, 202] = 0.22, p = 0.79, η2p = .00) and no
M. Mooijman, et al.
interaction effect between power and power stability, F(1, 202) = 2.23,
p = .14, η2p = .01. We observed no main effects of interaction effects on
power legitimacy, Fs < 1.39, ps > 0.28. These results confirm the
validity of our manipulations.
5.4.2. Trust
An ANOVA with power and power stability as independent vari-
ables and trust as dependent variable yielded a main effect of power, F
(1, 202) = 21.77, p < .001, η2p = .10. Participants in the high-power
condition (M = 3.14, SD = 1.57) trusted less than those in the low-
power condition (M = 4.14, SD = 1.56). Results also yielded a main
effect of power stability, F(1, 202) = 4.70, p = .031, η2p = .03.
Participants in the unstable condition (M = 3.40, SD = 1.64) trusted
less than those in the stable (M = 3.88, SD = 1.62). More important
and consistent with our predictions, results showed an interaction effect
between power and power stability, F(1, 202) = 5.11, p = .025,
η2p = .03. Trust was lower in the unstable high-power condition
(M = 2.67, SD = 1.40) as compared to the stable high-power condition
(M = 3.62, SD = 1.61; t[1 0 2] = 3.21, p = .002, d = 0.64), unstable
low-power condition (M = 4.15, SD = 1.54; t[1 0 0] = 5.13, p < .001,
d = 1.03) and stable low-power condition (M = 4.12, SD = 1.59), t
(1 0 2) = 4.94, p < .001, d = 0.97). Trust did not differ significantly
between the stable high-power condition and the stable low-power
condition (t[1 0 2] = 1.64, p = .11, d = 0.32; see Fig. 1). Trust did also
not differ significantly between the unstable low-power condition and
the stable low-power condition (t[1 0 0] = 0.12, p = 0.90, d = 0.02; see
Fig. 1).
5.5. Discussion
Study 1 supports the hypothesis that unstable (but not stable) power
decreases the extent to which individuals trust others. Legitimacy did
not explain this effect. In Study 2, we aimed to assess the general-
izability of these findings, by studying whether unstable power de-
creases the willingness to share resource with others (i.e., trust). Part of
the definition of trust is the willingness to open oneself up for potential
exploitation, for instance by making one’s resources available to others
(e.g., Berg et al., 1995).
5.6. Studies 2a and 2b
In studies 2a and 2b, we had two aims. First, as trusting others
entails the risk of making oneself vulnerable to potential exploitation
(i.e., managers trusting subordinates or other managers; Kramer, 1999),
we measured, using a Trust Game, in studies 2a and 2b the willingness
to share with others the resources that form the basis of one’s power.
This allowed us to extend our self-reported assessment of trust to the
measurement of trust behavior (Berg et al., 1995). Thus, an interaction
partner could gain control over participants’ valuable resources (and
gain more power relative to the participant). If concerns about losing
5
4.5
4 with the definition of trust; Johnson & Mislin, 2011; Pillutla, Malhotra,
Instability
Interpersonal Trust
3.5 Stability
3 ordinate—higher percentages of budget sent reflect higher trust. In
2.5
2 would give would be tripled and the other player could give back any
1.5
1
High-Power Low-Power
Fig. 1. Interpersonal trust as a function of power and stability for Study 1.
4
Organizational Behavior and Human Decision Processes 152 (2019) 1–10
power decrease trusting behavior, then we should observe a decrease in
trusting behavior when power is unstable. Second, in addition to trust
behavior, we also assessed economic risk-taking behavior in an In-
vestment Game. Instead of sharing resources with others, participants
could invest their resources without others potentially benefitting from
resource loss (such that resource loss would not be an interpersonal
power loss). If power-loss concerns are uniquely and negatively related
to trust, then we should not observe a decrease in risk-taking behavior
in the Investment Game when power is unstable; as losing resources is
not necessarily about losing power in this context. We ran two versions
of this paradigm. In Study 2a, participants imagined having the choice
to share their organizational budgets with an interaction partner and in
Study 2b, participants were given $10 and they could share this $10
with an interaction partner.
5.7. Method
5.7.1. Participants and design
In Study 2a, a total of 406 US college students (303 females;
Mage = 20.46 years, SDage = 2.14) were randomly assigned to a 2 (high
power vs. low power) × 2 (stable power vs. unstable power) between-
participants design, with trust vs. risk-taking as an additional within-
participants factor. In Study 2b, a total of 202 US college students (130
males; Mage = 20.82 years, SDage = 2.37) were randomly assigned to a
2 (high power vs. low power) × 2 (stable power vs. unstable power)
between-participants design, with trust vs. risk-taking as an additional
within-participants factor. For both studies, we aimed to recruit as
many participants as possible within a specific time frame (e.g.,
2 weeks). A power analysis using G*Power 7 shows that the combined
sample size of 608 participants provided 99% power to detect a
medium sized two-way interaction effect (f = 0.25).
5.8. Procedure
5.8.1. Power manipulation
For both studies, we used the same power manipulation as in Study
1.
5.8.2. Power stability manipulation
For both studies, we used the same power stability manipulation as
in Study 1.
5.8.3. Trust Game behavior
In Study 2a, consistent with previous work on the Trust Game (Berg
et al., 1995), participants were asked to imagine that they could send
any percentage of their budget to a subordinate who could then return
any amount of this budget to the participant. The budget participants
sent, however, would be tripled, whereas the money that would be sent
back would not. Playing by these rules, the participant sending 100% of
the budget would yield the subordinate 300% of the budget. The sub-
ordinate could then decide to send back to the participant any amount
of the tripled budget. Trust Games like these are frequently used to
measure behavioral displays of trust, because sending money to others
makes the participant vulnerable to potential exploitation (consistent
& Murnighan, 2003). To assess this behavior, participants were asked to
indicate what percentage of the budget they would send to the sub-
Study 2b, participants were told to imagine having $10 that they could
share with the other player: similar as in Study 2a, the money they
amount. Thus, the more money participants give in Study 2b, the more
they trust.
5.8.4. Investment Game behavior
Participants were also presented with an adapted version of the
M. Mooijman, et al. Organizational Behavior and Human Decision Processes 152 (2019) 1–10

Trust Game: the Investment Game. The order in which participants Table 1
completed the Trust Game and Investment Game was counterbalanced. Main effects and interaction effects for combined studies 2a and 2b.
In this version in Study 2a, the budget that participants sent to the df F p η2p
subordinate would still be tripled. It would be randomly determined,
however, which amount the participants would get back from the Trust/risk 1 (592) 1.14 0.29 0.00
Power 1 (592) 1.74 0.19 0.00
subordinate and the subordinate would not receive the budget that was
Power stability 1 (592) 12.55 < 0.001 0.02
“lost” by the participant. Thus, in Study 2a, the participant could invest Study 1 (592) 0.02 0.89 0.00
between 0 and 100% of budget, and receive back 300%. In Study 2b, Trust/risk × power 1 (592) 10.15 0.002 0.02
the same was the case except with participants being able to invest $10 Trust/risk × stability 1 (592) 4.94 0.027 0.01
and potentially getting back between $0 and $30. Power × stability 1 (592) 3.86 0.050 0.01
Trust/risk × study 1 (592) 2.74 0.098 0.01
Power × study 1 (592) 0.05 0.82 0.00
5.8.5. Manipulation checks Stability × study 1 (592) 13.47 < 0.001 0.02
We used the same item to measure perceived stability of one’s Trust/risk × power × stability 1 (592) 16.74 < 0.001 0.03
power position and the same three-item scale (α = 0.77; α = 0.92, Trust/risk × power × study 1 (592) 4.62 0.032 0.01
Trust/risk × stability × study 1 (592) 0.15 0.70 0.00
respectively) to measure perceived power as in Study 1.
Trust/risk × power × stability × study 1 (592) 0.33 0.56 0.00

5.9. Results
5.9.1. Manipulation checks
An ANOVA with power and power stability as independent vari-
ables and perceived power as dependent variable yielded a main effect
of power, F(1, 602) = 60.12, p < .001, η2p = .10. Participants in the
high-power condition (M = 4.14, SD = 1.73) felt more powerful than
participants in the low-power condition (M = 3.15, SD = 1.39). We
observed no main effect of power stability, F(1, 602) = 0.95, p = 0.76,
η2p = .00, and no interaction effect between power and power stability
on perceived power, F(1, 602) = 0.32, p = .57, η2p = .00.
In addition, an ANOVA with power and power stability as in-
dependent variables and perceived stability as dependent variable
yielded a main effect of power stability, F(1, 604) = 39.44, p < .001,
η2p = .06. Participants in the unstable power condition (M = 4.74,
SD = 1.34) felt more unstable in their power than participants in the
stable condition (M = 3.96, SD = 1.70). We observed no main effect of
power, F(1, 604) = 0.31, p = .58, η2p = .00, and no interaction effect
between power and power stability on perceived power stability, F(1,
604) = 1.50, p = .22, η2p = .01.
5.9.2. Trust Game and Investment Game behaviors
We combined both studies into one analysis, recoding the amount
between $0 and $10 given to the other player by participants in Study
2a as between 0% and 100% of their “budget”. We used a repeated-
measures ANOVA with power, stability, and study (i.e., Study 2a or 2b)
as an additional factor. This repeated-measures ANOVA with power,
stability, and study as independent variables and trust and risk-taking
as dependent variables, yielded the predicted three-way interaction
between power, stability, and the within-factor of trust or risk-taking, F
(1, 592) = 16.74, p < .001, η2p = .03. The analysis did not yield a
four-way interaction between power, stability, trust vs. risk-taking and
study type (F[1, 592] = 0.33, p = .56, η2p = .00), showing that study
type did not moderate the observed effects. See Table 1 for all other
main effects and interaction effects.
For our follow-up analyses, we first conducted an ANOVA with
power and stability as independent variables and trust as dependent
variable. Overall, participants in the high-power conditions
(M = 60.53, SD = 15.14) trusted less than those in the low-power
conditions (M = 65.18, SD = 15.17; F[1, 604] = 14.89, p < .001,
η2p = .03) and participants in the unstable conditions (M = 60.81,
SD = 14.91) trusted less than those in the stable conditions
(M = 64.90, SD = 15.47; F[1, 604] = 11.69, p = .001, η2p = .02). In
addition, an interaction effect between power and stability on trust (F
[1, 604] = 18.41, p < .001, η2p = .03) demonstrated that trust was
lower in the unstable high-power condition (M = 55.90, SD = 12.39) as
compared to the stable high-power condition (M = 65.17, SD = 16.22;
t[3 0 0] = 5.58, p < .001, d = 0.65), unstable low-power condition
(M = 65.70, SD = 15.63; t[3 0 1] = 6.05, p < .001, d = 0.69), and
stable low-power condition (M = 64.66, SD = 14.73), t[3 0 3] = 5.61,
p < .001, d = 0.65. Trust did not differ significantly between the
stable high-power condition and the stable low-power condition, t
(3 0 1) = 0.29, p = .77, d = 0.03, and between the stable high-power
condition and the unstable low-power condition, t(3 0 3) = 0.29,
p = .77, d = 0.03. Trust did not differ between the stable low-power
condition and the unstable low-power condition, t(3 0 4) = 0.60,
p = .55, d = 0.07.
Second, we conducted an ANOVA with power and stability as in-
dependent variables and economic risk-taking as dependent variable.
Participants in the high-power condition (M = 64.56, SD = 14.59) took
slightly more economic risks than those in the low-power condition
(M = 62.28, SD = 12.56), F(1, 602) = 4.29, p = .039, η2p = .01.
Participants in the unstable condition took as many risks (M = 63.32,
SD = 13.88) as those in the stable condition (M = 63.39, SD = 13.41),
F(1, 602) = 0.02, p = .90, η2p = .00. An interaction effect between
power and stability on trust (F[1, 602] = 4.67p = .031, η2p = .01) was
observed but with only one significant contrast: participants took more
economic risks when comparing the unstable high-power condition
(M = 65.69, SD = 15.05) to the unstable low-power condition
(M = 61.01, SD = 12.25), t(2 9 6) = 2.96, p = .003, d = 0.34. Risk-
taking did not significantly differ between the other conditions (stable
high-power position; M = 65.17, SD = 16.22; stable low-power posi-
tion; M = 64.66, SD = 14.73, all contrasts ts < 1.76, ps > 0.10,
d < 0.17).
5.10. Discussion
These findings replicated the results from Study 1 in a different
setting. Unstable power decreased the willingness to give away control
over one’s resources to others (interpersonal trust) but it—although
somewhat unreliably—increased the willingness to take economic risks
in the Investment Game (consistent with findings from Jordan et al.,
2011). These findings provide additional evidence for the notion that
unstable power decreases trust because it raises concerns about losing
power to others. Yet, in studies 1 and 2, we assumed that unstable
power raises power-loss concerns. In studies 3 and 4, we directly tested
this assumption by explicitly assessing power-loss concerns.
5.11. Study 3
Study 3 had two goals. First, in addition to assessing expectations
about others to indicate trust (cf. Study 1), in this study we also ex-
amined trust displayed in behavior towards these others (cf. Study 2).
We aimed to demonstrate that the power-loss concerns that are raised
by the lack of power stability impact Trust Game behavior through
decreasing benevolent trust expectations (cf. studies 1 and 2). Second,
we compared stable and unstable high-power conditions to a high-
power condition in which no information was given about power

5
M. Mooijman, et al.
stability. This allowed us to investigate whether the stable or unstable
high-power conditions deviate from the default effects of power on trust
(see also Brion, Mo, & Lount, 2019).
5.12. Method
5.12.1. Participants and design
A total of 156 Mechanical Turk participants (97 males;
Mage = 33.56 years, SDage = 8.53) participated for $1 and were ran-
domly assigned to one of three high-power conditions: stable high-
power, unstable high-power, or control. A power analysis using
G*Power 7 shows that this sample size provided 99% power to detect
the smallest effect size that we observed in Studies 1 and 2 (d = 0.65).
5.13. Procedure
5.13.1. Power stability manipulation
Participants were informed that they would be working together on
a task with another participant, either as leader or subordinate.
Participants were informed that the leader could evaluate the sub-
ordinate at the end of the experiment and that this evaluation would
influence the amount of money the subordinate would get for partici-
pating in the experiment (Lammers et al., 2008; Mooijman et al., 2015).
Participants were also informed that the subordinate did not have this
opportunity. All participants filled out a ten-item “Leadership Assess-
ment Questionnaire”, allegedly to determine whether they would be
assigned the role of leader or subordinate (e.g., to what extent do you
agree with the statement: “A leader should always seek input from
others”). Although participants were not given feedback on their scores,
all participants were informed that, on the basis of their scores, they
were assigned to the leadership position (cf. Lount & Petitt, 2012;
Mooijman et al., 2015; Stouten, De Cremer, & Van Dijk, 2005).
Participants in the control condition were provided with no further
information. Participants in the unstable high-power condition were
informed that position could change during the experiment, such that
they might end up being a subordinate. Participants in the stable high-
power condition were informed that their position could not change
during the experiment.
5.13.2. Trust Game behavior
Participants were given the same instructions about the Trust Game
as in Study 2b. We then also assessed their expectations of the trust-
worthiness of the subordinate with a seven-point, three-item scale: “I
think the subordinate cannot be fully trusted to give the money back to
me”, “I distrust the subordinate a little bit”, and “I think the subordinate
is inclined to keep the money for him/herself” (α = 0.83).
5.13.3. Power-loss concerns
Power-loss concerns were measured with a three-item scale: “I fear
losing my power position”, “I dread the possibility of being a sub-
ordinate”, and “I don’t like losing my position” (α = 0.94).
5.13.4. Manipulation checks
Perceived stability of power was measured with a three-item scale:
“I might lose my leadership position”, “I may become a subordinate”,
and “My leadership position is somewhat unstable” (α = 0.98; higher
scores indicated more power instability). To check whether the power
stability manipulation affected participants' sense of power, we also
measured perceived power with a three-item scale: “I feel powerful”, “I
occupy a position of power”, and “I have power over the subordinate”
(α = 0.88).
5.14. Results
5.14.1. Manipulation checks
An ANOVA with the three high-power conditions as independent
6
Organizational Behavior and Human Decision Processes 152 (2019) 1–10
variable and perceived stability of power as dependent variable yielded
a main effect of our power manipulation, F(2, 153) = 61.63, p < .001,
η2p = .45. Participants perceived their power as more unstable in the
unstable high-power condition (M = 5.83, SD = 1.54) compared to the
stable high-power condition (M = 2.16, SD = 1.85; t[1 0 1] = 10.92,
p < .001, d = 2.17) as well as the control condition (M = 3.17,
SD = 1.77; t[1 0 1] = 8.13, p < .001, d = 1.62). Furthermore, parti-
cipants in the stable high-power condition perceived their position as
less unstable than participants in the control condition, t(1 0 4) = 2.88,
p = .005, d = 0.56. These results confirm the validity of our power
manipulation. Additional checks indicated, as intended, that partici-
pants in the stable high-power condition (M = 5.43, SD = 1.43), un-
stable high-power condition (M = 5.21, SD = 1.35), and control con-
dition (M = 5.34, SD = 1.14) felt equally powerful, F(2, 153) = 0.38,
p = .69, η2p = .00; ts < 1, ps > 0.43, ds < 0.15.
5.14.2. Power-loss concerns
An ANOVA with high-power condition as independent variable and
power-loss concerns as dependent variable yielded a main effect of
condition, F(2, 153) = 7.50, p = .001, η2p = .09. Power-loss concerns
were stronger in the unstable high-power condition (M = 3.66,
SD = 1.94) compared to the stable high-power condition (M = 2.38,
SD = 1.83; t[1 0 1] = 3.46, p = .001, d = 0.69) and high-power control
condition (M = 2.60, SD = 1.58; t[1 0 1] = 3.04, p = .003, d = 0.60).
Power-loss concerns did not differ between the stable high-power
condition and the control condition, t(1 0 4) = 0.68, p = .49, d = 0.13.
5.14.3. Trust Game behavior
An ANOVA with power condition as independent variable and trust
as dependent variable yielded a main effect of condition, F(2,
151) = 3.13, p = .047, η2p = .04. Trust was lower in the unstable high-
power condition (M = 4.25, SD = 3.08) compared to the stable high-
power condition (M = 5.81, SD = 3.55; t[1 0 1] = 2.36, p = .020,
d = 0.047) and control condition (M = 5.68, SD = 3.77; t
[1 0 1] = 2.09, p = .038, d = 0.042). Trust did not differ between the
stable high-power condition and the control condition, t(1 0 1) = 0.17,
p = .86, d = 0.00.
A similar analysis with expectations of the subordinate’s perceived
trustworthiness as the dependent variable also yielded a main effect of
condition, F(2, 153) = 3.16, p = .045, η2p = .04. Participants in the
unstable high-power condition expected more untrustworthy behavior
(M = 5.17, SD = 1.37) than participants in the stable high-power
condition (M = 4.45, SD = 1.73; t[1 0 1] = 2.28, p = .025, d = 0.45)
and control condition (M = 4.47, SD = 1.78; t[1 0 1] = 2.27, p = .025,
d = 0.45). Expectations of trustworthiness did not differ between the
stable high-power condition and control condition, t(1 0 4) = 0.04,
p = .97, d = 0.00.
5.14.4. Mediation analysis
Confirming our reasoning that the effect of power stability on the
amount of money sent in the Trust Game should be due to decreased
expectations of trustworthiness when power is unstable, a boot-
strapping analysis using 10,000 resamples (Hayes, Preacher, & Myers,
2011) demonstrated that trustful expectations mediated the effect of
the power position stability manipulation on the amount of money sent,
95% CI = [−0.72, −0.13]. If the decreased trust in the unstable high-
power condition was due to an increase in power-loss concerns, then
this should mediate the effect of power stability on expectations of
trustworthiness predicting Trust Game behavior (power stability
power-loss concerns expectations of trustworthiness Trust Game beha-
vior). We tested this multiple-step model using a macro developed by
Hayes et al. (2011). As expected, power-loss concerns correlated ne-
gatively with the amount of money sent in the Trust Game (r = −0.31,
p < .001) and correlated negatively with having trust expectations
(r = −0.29, p < .001). Moreover, the combined effect of power-loss
concerns through expectations of trustworthiness on Trust Game
M. Mooijman, et al.
behavior was significant, 95% CI = [−0.35, −0.06].
5.15. Discussion
Study 3 supports the hypothesis that unstable power decreases trust,
and that this impacts interpersonal behavior. Specifically, occupying an
unstable power position increased concerns about power loss, which, in
turn, undermined the expected trustworthiness of others, as well as
trusting behavior towards others. Importantly, trust in the unstable
high-power condition was lower than in the control condition, but the
stable high-power position did not increase trust compared to the
control condition. This is consistent with the finding that the crucial
underlying variable (i.e., power-loss concerns) did not differ between
the stable high-power condition and the control condition (participants
in both conditions were relatively unconcerned about losing their
power), whereas power-loss concerns were increased for participants in
the unstable high-power condition. In addition, these findings further
corroborate and extend the findings from studies 1 and 2, showing that
power did not undermine trust when power was stable. Study 3 thus
provides further support for the argument that power-loss concerns
explain (in part) why power instability undermines trust. The results
also suggest, however, that the default response to high power is not to
raise concerns with position loss or distrust. Instead, these responses to
high power only seem to emerge when people are explicitly reminded
that their position might be unstable.
5.16. Study 4
So far, all our previous studies have been experimental studies. In
Study 4, we aimed to demonstrate the relation of unstable power with
trust in a field study, using organizationally relevant measures.
Specifically, we measured the degree to which managers’ control over
valuable resources in their organization, and the perceived stability of
this resource control, were related to their trust towards others in the
organization.
5.17. Method
5.17.1. Participants and design
We recruited working professionals. They were informed that we
conducted a study on everyday work experiences and that we were
interested in exploring the relationships between managers and their
subordinates. We thus asked for manager-employee pairs who were
employed full-time to participate in our study. When interested, they
were directed to provide us with contact information (work e-mail) for
their subordinate (when they were a manager) or manager (when they
were a subordinate). A total of 96 supervisors (58 females;
Mage = 40.51 years, SDage = 10.35) with a corresponding subordinate
(56 females; Mage = 35.40 years, SDage = 12.08) were successfully re-
cruited. Our sample size was determined, and constrained by, research
funds; we aimed to recruit 100 supervisor-subordinate pairs. All su-
pervisors worked for (a variety of) moderate to large firms at the time of
the survey. The supervisors had been with the organization, on average,
for 8.91 years (SD = 7.93) and in their position, on average, for
5.69 years (SD = 6.83). The survey contained scales measuring super-
visors’ sense of power and perceived power stability, and included
several demographic questions. Subordinates were asked to rate the
extent to which their corresponding manager tended to trust others in
the organization. We used such supervisor-subordinate dyads to elim-
inate common method bias and to be able to establish the link between
power, stability, and trust in situations with real-life power and trust
differences.
7
Organizational Behavior and Human Decision Processes 152 (2019) 1–10
5.18. Procedure
5.18.1. Power measurement
Consistent with the notion that power entails control over scarce
resources, supervisors indicated on a seven-point scale the degree to
which they had personal control over (1) organizational resources, (2)
budgets, and (3) flows of information (1 = very little control, 7 = a lot of
control; α = 0.95).
5.18.2. Power stability
Supervisors indicated on a three-item scale to what extent they
perceived their control over resources, budgets, and information to be
stable or unstable (1 = very unstable, 7 = very stable; α = 0.85).
5.18.3. Power-loss concerns
Supervisors indicated on a three-item scale to what extent they were
concerned about losing control over resources, budgets, and informa-
tion (1 = not concerned at all, 7 = very concerned; α = 0.91).
5.18.4. Trust
Consistent with the notion that trust entails the willingness to be
vulnerable to others, in particular the willingness to share resources
with others (Berg et al., 1995; Kramer, 1999), the subordinates in-
dicated on a three-item scale to what extent their manager was willing
to share organizational resources, budgets, and information with them
(1 = very unwilling, 7 = very willing; α = 0.75).3
5.19. Results
See Table 2 for table with relevant correlations.
5.19.1. Power-loss concerns
Regression analyses were used to test the interactive effect of power
and stability with power-loss concerns. For the first step, power and
stability were added. For the second step, the interaction between
power and power stability was added. Results demonstrated that more
power was associated with higher power-loss concerns (β = 0.24, t
[96] = 2.43, p = .006) and that stability was, overall, negatively re-
lated to power-loss concerns (β = −0.17, t[96] = 1.71, p = .091).
Crucially, we observed an interaction between power and stability,
β = −0.44, t[96] = 3.83, p < .001. Power was stronger related to
power-loss concerns when power was unstable (+1 SD; β = 0.36, t
[96] = 2.89, p = .005) compared to when power was stable (−1 SD;
β = 0.06, t[96] = 0.44, p = .66).
5.19.2. Trust
Results from a similar regression with trust as dependent variable
demonstrated that power was negatively related to trust behavior
(β = −0.25, t[96] = −2.49, p = .014) but that stability was unrelated
to trust behavior (β = −0.04, t[96] = −0.43, p = .66). Crucially, we
observed an interaction between power and stability, β = 0.22, t
(96) = 2.65, p = .009. Power decreased trusting behavior when power
was unstable (+1 SD; β = −0.43, t[96] = −3.37, p = .001) but not
when power was stable (−1 SD; β = .−01, t[98] = −0.05, p = .97).
5.19.3. Mediation analysis
Replicating the previous three studies, power-loss concerns were
negatively correlated with trust, r = −0.29, p = .003. We used Model 8
with 5000 bootstraps in PROCESS (Hayes et al., 2011) to test the degree
to which power-loss concerns mediated the relationship between power
3
We also added a scale that directly measured trust expectations (e.g., my
supervisor trusts me; 1 = completely disagree; 7 = completely agree; α = 0.92).
Unstable power also undermined this measure of trust and mediated the impact
of unstable power on the behavioral indicator of trust (95% CI [-0.64, −0.13]).
M. Mooijman, et al.
Table 2
Correlation table for Study 4.
Power Power stability Power-loss concerns Trust
Power – evidence as to the conditions under which this does (not) happen. The
Power stability 0.01 –
Power-loss concerns 0.23* −0.20* –
Trust −0.25* 0.19 −0.30** –
Note. ***p < .001.
* p < .05
** p < .01.
and trust expectations and trust behavior when we allow stability to
moderate the relationship between power and trust (as demonstrated
earlier). Results yielded a significant moderated mediation effect, 95%
CI = [0.02, 0.17]. Power-loss concerns mediated the negative re-
lationship between power and trust only when power was unstable (−1
SD; 95% CI = [−0.31, −0.03]), but not when power was stable (1 SD;
95% CI = [−0.03, 0.11]).4
5.20. Discussion
Study 4 replicated in a field study the findings from the previous
three experimental studies and demonstrates the impact of unstable
power on trust in a supervisor-subordinate dyadic sample. These find-
ings provide additional evidence for the notion that occupying an un-
stable power position decreases a power holder’s trust through raising
power-loss concerns, although it is also possible in Study 4 that man-
agers with unstable power are less willing to share resources because
they think it’s not their place to share those resources with others
(because someone else might take over their position). We discuss this
possibility further in the next section.
6. General discussion
Trusting others is vital for leaders. In the present research, we ex-
amined conditions under which power decreases trust and the process
by which this occurs. We demonstrated in four studies that making
salient the lack of stability in high-power positions decreases trust as it
raises individuals’ concerns about losing power. By contrast, occupying
a stable power position attenuates the extent to which power decreases
trust as it alleviates individuals’ concerns about losing power. Indeed,
unstable power decreased trust regardless of whether we provided
participants with a justification for their unstable power position—the
results of Study 3 (where we did not provide a reason for why parti-
cipants’ power was unstable) matched the findings of Studies 1 and 2 in
which we provided an explicit rationale for why power was unstable
(i.e., the firm wants managers to change divisions in order to stimulate
productivity). Moreover, the fact that we replicated the effect of un-
stable power on trust in a field study of managers and subordinates
(Study 4), suggests that the effect of unstable power on trust is not just
an experimental methodological artifact. Together, these studies pro-
vide converging support for the hypothesis that emphasizing the in-
stability of high-power positions undermines trust in other people’s
good intentions and reduces the willingness to engage in trusting be-
havior. These findings are important as they add insights to the lit-
erature on power, trust, and risk-taking, as detailed next.
6.1. Implications and Future directions
The present research makes several contributions to the literature
4
Please note that we also measured supervisor power from the perspective of
the subordinate and found similar effects as reported in the results section of
Study 4.
8
Organizational Behavior and Human Decision Processes 152 (2019) 1–10
on power and trust. First, it elucidates when and why power under-
mines trust. Previous research has demonstrated that having power can
decrease an individual’s trust in others (Kramer & Gavrieli, 2004;
Mooijman et al., 2015; Schilke et al., 2015), but has provided little
current research indicates that power decreases trust when power-loss
concerns are salient due to a lack of power stability, but this is not a
default response, as power does not always decrease trust when power
is stable and power-loss concerns are not salient, compared to high-
power control conditions. This is consistent with the proposed argu-
ment that people are reluctant to make themselves vulnerable to others
when there is a potential danger of losing one’s power. Indeed, since
many leader-follower interactions take place in a hierarchical context,
with leaders (e.g., managers or governmental officials) having power
over others (e.g., employees or citizens), this suggests that low trust
proliferates when leaders are reminded of the possibility of losing their
power to others.
It seems common practice for organizations to reposition their
employees from one division to another in order to reduce costs and to
survive in a continuously transforming environment (Amabile & Conti,
1999). Although this is commonly assumed to stimulate productivity
through making employees expend more effort and reflect on their own
practices, we demonstrate that this may also undermine the extent to
which power holders trust others in the organization. Indeed, while the
threat of position loss is often used as an incentive in organiza-
tions—intended to increase motivation and performance of those in
leadership roles—this research documents the potentially adverse ef-
fects of this motivation strategy, as it undermines trust. Interestingly,
recent research has demonstrated that feeling distrusted by leaders
undermines people’s motivation to stick to the rules, thereby stimu-
lating unethical behavior (Mooijman et al., 2017). The current research
thus identifies power instability as a relevant moderator of high-power
effects, as it raises power-loss concerns. Specifically, our findings de-
monstrate that having power can be associated with a concern for
power loss and that this, in turn, undermines trust (even though this
effect is attenuated when power is stable).
Future research might investigate additional moderators of high-
power effects and alternative concerns that might be relevant to trust
and cooperative behavior in organizations. These might include in-
dividual difference variables (e.g., power motivation) as well as other
situational characteristics that define the power role (e.g., account-
ability, feelings of responsibility, and legitimacy; Tost, 2015). Our re-
sults (in particular the additional checks in Study 1) demonstrate that
the effect of power stability can be distinguished from power legiti-
macy. Legitimacy, however, tends to moderate the effects of power,
with illegitimate power making individuals more instead of less in-
hibited (Lammers et al., 2008) and less instead of more corrupt
(Lammers, Stapel, & Galinsky, 2010). An interesting avenue for future
research might be to investigate the impact of power legitimacy on
interpersonal trust. Possibly, power decreases trust especially when
power is illegitimate. Illegitimate power hierarchies are characterized
by forced compliance by power holders and resistance from the less
powerful (Lenski, 1966; Mills, 1956). Indeed, power holders may an-
xiously try to protect their positions of power in such hierarchies (and
become inhibited instead of action-oriented; Lammers et al., 2008),
possibly lowering their trust in others. In addition, previous research
has suggested that unstable power can increase non-social (economic)
forms of risk-taking (Jordan et al., 2011). Yet, our research suggests
that power-loss concerns play an important role when resources can be
lost that form the basis of one’s power and that this can decrease the
degree to which power holders become trusting. Our findings thus show
that unstable power does not always foster social forms of risk-taking.
Second, the current research broadens the understanding of the
determinants of trust. Trust is a key variable in social interactions, and
has been a much-studied topic in social hierarchies. Nevertheless, few
studies to date have demonstrated empirically when and why power
M. Mooijman, et al.
differences within social hierarchies affect interpersonal trust and dis-
trust. We demonstrate a psychological mechanism that can explain
when and why power affects trust. Given the influence of leaders in
setting exemplary behavior (Mayer, Kuenzi, Greenbaum, Bardes, &
Salvador, 2009) and the negative effects associated with low levels of
trust (e.g., decreased liking, increased tendency for unethical behavior;
Mulder et al., 2006), any condition that lowers trust is potentially
problematic as it communicates and spreads low trust within an orga-
nization or society. Indeed, initial trust between people tends to be
considerable (see Kramer & Lewicki, 2010, for a review), but unstable
power hierarchies seem to undermine this trust—at least from the part
of the power holder. This can have important implications for attempts
to promote interpersonal trust in organizations. These attempts may fail
when organizations reposition employees and create positional in-
stability. This is consistent with previous research that has suggested
that unstable work group membership makes it less likely that people
value and accept constructive task contributions offered by newcomers
(Rink & Ellemers, 2015). As workgroup instability can undermine (ra-
ther than foster) change and innovation, instability of power positions
can undermine (rather than foster) trust in work teams. Promoting in-
terpersonal trust may be especially effective when organizations pro-
vide a stable basis of power.
Trust is likely to decrease when power-loss concerns are salient,
whereas trust may increase when leaders feel like they have status
within the hierarchy (i.e., feel respected and admired; Blader & Chen,
2012; Lount & Petitt, 2012). Making leaders feel respected and admired
indeed has the potential to offset the negative effects of leaders’ power
on trust (Lount & Petitt, 2012). Given that both power and status dif-
ferences are ever present in organizational hierarchies, we argue it
would be fruitful to investigate how leaders’ power and status interact
to influence trust in others, and to investigate to what extent sub-
ordinates trust leaders with power and/or status. For instance, due to
the status-conferral process, status tends to focus attention outward,
increasing the realization that others are sensitive to the treatment they
receive and increasing the expectation that those who confer status on
you will treat you well (e.g., trusting behavior; Blader & Chen, 2012;
Blader, Shirako, & Chen, 2016). In contrast, the impact of power on
trust may depend on how individuals have gained their power and how
they construe their power. For instance, power can be construed as
providing an individual with increased opportunity to pursue personal
goals or with increased responsibility for others (e.g., De Wit,
Scheepers, Ellemers, Sassenberg, & Scholl, 2017; Sassenberg, Ellemers,
Scheepers, & Scholl, 2014). When power is construed as providing the
opportunity to fulfill personal goals, it may draw an individual’s at-
tention inwards towards personal goals (Galinsky, Magee, Gruenfeld,
Whitson, & Liljenquist, 2008; Guinote, 2017), with concerns about
personal gain and potential loss weighing heavier than others’ needs
(Blader & Chen, 2012). Since trusting others can lead to direct loss over
resource-control, whereas trusting others is more likely to increase re-
spect from others (Tyler, Jackson, & Mentovich, 2015), power con-
strued as opportunity may have a different effect on trust than status,
especially when power is unstable. When, however, power is construed
as responsibility for others (e.g., managers that are responsible for
subordinates’ needs), it may draw an individual’s attention outward
towards others’ needs and concerns, thereby possibly fostering trust.
Following this line of reasoning, gaining power by taking it away from
others may highlight the personal opportunities of power (thereby de-
creasing trust), whereas being granted power by others may highlight
the responsibilities of power (thereby increasing trust). Future research
could investigate these interesting propositions further.
It is worth noting that the set-up of the current research somewhat
restricts the generalizability of our findings. That is, other variables
may also play a role in organizations. For instance, we have oper-
ationalized trust mainly as expectations about the (benevolent or mal-
icious) intentions of others, but power could also affect trust as an ex-
pectation of competency instead of integrity (Mayer et al., 1995). It is
9
Organizational Behavior and Human Decision Processes 152 (2019) 1–10
unclear how power would affect such competency-based trust, although
power holders’ illusion of control or belief in meritocracy may increase
the belief that others’ work is in part due to the power holder (Fast,
Gruenfeld, Sivanathan, & Galinsky, 2009), thereby decreasing percep-
tions of competency-based trust. Moreover, in the current studies,
power holders were relatively independent from others (cf. the defini-
tion of power; Magee & Galinsky, 2008). In many instances of real life,
power holders are also dependent on their subordinates (e.g., politi-
cians relying on citizens’ votes). Future research could investigate how
this form of dependence interacts with leaders’ power-loss concerns to
predict low trust towards subordinates. Being dependent on sub-
ordinates for maintaining one’s power may, for instance, foster trust
towards subordinates, especially for those individuals who are highly
motivated to maintain power—because the only way to maintain power
is through their subordinates. Future research could investigate these
issues further. Lastly, we might have observed some of the trust beha-
viors in the current research because power holders may think they do
not have the legitimacy to share resources when their power is unstable
(a potential newcomer might be given that right). That said, studies 1
and 3 explicitly measured trust expectations instead of sharing beha-
vior; we also observed the effects of unstable power on trust expecta-
tions in those studies. In addition, trust expectations mediated the effect
of unstable power on sharing behavior in our studies, and feelings of
legitimacy did not mediate the effects of unstable power on trust in
Study 1. As such, we are confident that we provide convergent evidence
for the idea that unstable power reduces trust and that the observed
reduction in sharing behavior is at least partly due to a reduction in
trust (and not completely due to legitimacy concerns), consistent with
our theorizing.
7. Conclusion
We presented four studies examining when and why power under-
mines trust and how this can be prevented. Using different manipula-
tions of power, different samples of participants, and different in-
dicators of trust, we consistently observed that trust depended on
leaders’ concerns about losing power, raised by the instability of high-
power. Power-loss concerns can thus explain when power decreases
trust in others and how—through decreasing these concerns—this can
be attenuated. In doing so, the current work broadens knowledge about
the effects of power and the determinants of trust. It also suggests
practical implications for managers, leaders, and policy-makers about
why their power can make them less trusting and how to prevent this.
Appendix A. Supplementary material
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.obhdp.2019.03.009.
References
Amabile, T. M., & Conti, R. (1999). Changes in the work environment for creativity during
downsizing. Academy of Management Journal, 42, 630–640.
Anderson, C., & Brion, S. (2014). Perspectives on power in organizations. Annual review of
Organizational Psychology and Organizational Behavior, 1, 67–97.
Berg, J., Dickaut, J., & McCabe, K. (1995). Trust, reciprocity, and social history. Games
and Economic Behavior, 10, 122–142.
Berger, J., Rosenholtz, S. J., & Zelditch, M. (1980). Status organizing processes. Annual
Review of Sociology, 6, 479–508.
Blader, S., & Chen, Y. (2012). Differentiating the effects of status and power: A justice
perspective. Journal of Personality and Social Psychology, 102, 994–1014.
Blader, S. L., Shirako, A., & Chen, Y. (2016). Looking out from the top: Differential effects
of status and power on perspective taking. Personality and Social Psychology Bulletin,
42, 723–737.
Bombari, D., Schmid, Mast M., & Bachmann, M. (2017). Felt power explains the link
between position power and experienced emotions. Emotion, 17, 55–66.
Brion, S., Mo, R., & Lount, R. B. (2019). Dynamic influences of power on trust: Changes in
power affects trust in others. Journal of Trust Research (in press).
Brower, H. H., Lester, S. W., Korsgaard, M. A., & Dineen, B. R. (2009). A closer look at
trust between managers and subordinates: Understanding the effects of both trusting
M. Mooijman, et al.
and being trusted on subordinate outcomes. Journal of Management, 35, 327–347.
Brown, S., Gray, D., McHardy, J., & Taylor, K. (2015). Employee trust and workplace
performance. Journal of Economic Behavior & Organization, 116, 361–378.
Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s Mechanical Turk a new
source of inexpensive, yet high-quality, data? Perspectives on Psychological Science,
6, 3–5.
Buhrmester, M. D., Talaifar, S., & Gosling, S. (2018). An evaluation Amazon’s Mechanical
Turk, its rapid rise, and its effective use. Perspectives on Psychological Science, 13,
149–154.
Case, C. R., & Maner, J. K. (2014). Divide and conquer: When and why leaders undermine
the cohesive fabric of their group. Journal of Personality and Social Psychology, 107,
1033–1050.
De Jong, S. B., Van der Vegt, G. S., & Molleman, E. (2007). The relationships among
asymmetry in task dependence, perceived helping behavior, and trust. Journal of
Applied Psychology, 6, 1625–1637.
De Wit, F. R. C., Scheepers, D. T., Ellemers, N., Sassenberg, K., & Scholl, A. (2017).
Whether power holders construe their power as responsibility or opportunity influ-
ences their tendency to take advice from others. Journal of Organizational Behavior,
38(7), 923–949.
Fast, N. J., Gruenfeld, D. H., Sivanathan, N., & Galinsky, A. D. (2009). Illusory control: A
generative force behind power’s far-reaching effects. Psychological Science, 20,
502–508.
Fehr, E., Herz, H., & Wilkening, T. (2013). The lure of authority: Motivation and incentive
effects of power. American Economic Review, 103, 1325–1359.
Feld, L. P., & Frey, B. S. (2002). Trust breeds trust: How taxpayers are treated. Economics
of Governance, 3, 87–99.
Fetchenhauer, D., & Dunning, D. (2012). Betrayal aversion versus principled
trustfulness—How to explain risk avoidance and risky choices in trust games. Journal
of Economic Behavior and Organization, 81, 534–541.
Galinsky, A. D., Gruenfeld, D. H., & Magee, J. C. (2003). From power to action. Journal of
Personality and Social Psychology, 85, 453–466.
Galinsky, A. D., Magee, J. C., Gruenfeld, D. H., Whitson, J., & Liljenquist, K. A. (2008).
Social power reduces the strength of the situation: Implications for creativity, con-
formity, and dissonance. Journal of Personality and Social Psychology, 95, 1450–1466.
Goodwin, S. A., Gubin, A., Fiske, S. T., & Yzerbyt, V. Y. (2000). Power can bias impression
processes: Stereotyping subordinates by default and by design. Group Processes &
Intergroup Relations, 3, 227–256.
Guinote, A. (2007). Power and goal pursuit. Personality and Social Psychology Bulletin, 33,
1076–1087.
Guinote, A. (2017). How power affects people: Activating, wanting, and goal seeking.
Annual Review of Psychology, 68, 353–381.
Hauser, D. J., & Schwarz, N. (2016). Attentive turkers: MTurk participants perform better
on online attention checks than subject pool participants. Behavior Research Methods,
48, 400–407.
Hayes, A. F., Preacher, K. J., & Myers, T. A. (2011). Mediation and the estimation of
indirect effects in political communication research. In E. P. Bucy, & R. Lance Holbert
(Eds.). Sourcebook for political communication research: Methods, measures, and analy-
tical techniques (pp. 434–465). New York: Routledge.
Houser, D., Schunk, D., & Winter, J. (2010). Distinguishing trust from risk: An anatomy of
the trust game. Journal of Economic Behavior and Organization, 74, 71–81.
Inesi, E. M., Gruenfeld, D. H., & Galinsky, A. D. (2012). How power corrupts relationships:
Cynical attributions for others’ generous acts. Journal of Experimental Social
Psychology, 48, 795–803.
Johnson, N. D., & Mislin, A. A. (2011). Trust games: A meta-analysis. Journal of Economic
Psychology, 32, 865–889.
Jordan, J., Sivanathan, N., & Galinsky, A. D. (2011). Something to lose and nothing to
gain: The role of stress in the interactive effect of power and stability on risk taking.
Administrative Science Quarterly, 56, 530–558.
Keltner, D., Anderson, C., & Gruenfeld, D. H. (2003). Power, approach, and inhibition.
Psychological Review, 23, 265–284.
Kirchler, E., Kogler, C., & Muelbacher, S. (2014). Cooperative tax compliance: From de-
terrence to deference. Current Directions in Psychological Science, 23, 87–92.
Kramer, R. M. (1999). Trust and distrust in organizations: Emerging perspectives, en-
during questions. Annual Review of Psychology, 50, 984–993.
Kramer, R. M., & Gavrieli, D. A. (2004). Power, uncertainty, and the amplification of
doubt: Distrust and suspicion in the oval office. In R. M. Kramer, & K. S. Cook (Eds.).
Trust in organizations: Dilemmas and approaches (Russell Sage Foundation Trust Series,
Volume VI) (pp. 342–370). New York: Russell Sage Foundation.
Kramer, R. M., & Lewicki, R. (2010). Repairing and enhancing trust: Approaches to re-
ducing organizational trust deficits. Academy of Management Annals, 4, 245–277.
Lammers, J., Galinsky, A. D., Gordijn, E. H., & Otten, S. (2008). Illegitimacy moderates
the effects of power on approach. Psychological Science, 19, 558–564.
Lammers, J., Stapel, D. A., & Galinsky, A. D. (2010). Power increases hypocrisy:
Moralizing in reasoning, immorality in behavior. Psychological Science, 21, 737–744.
10
Organizational Behavior and Human Decision Processes 152 (2019) 1–10
Lammers, J., Stoker, J. I., Jordan, J., Pollmann, M. M. H., & Stapel, D. A. (2011). Power
increases infidelity among men and women. Psychological Science, 22, 1191–1197.
Lau, D. C., & Lam, L. W. (2008). Effects of trust and being trusted on team citizenship
behavior in chain stores. Asian Journal of Social Psychology, 11, 141–149.
Lenski, G. E. (1966). Power and privilege: A theory of social stratification. New York:
McGraw-Hill.
Lewicki, R. J., Tomlinson, E. C., & Gillespie, N. (2006). Models of interpersonal trust
development: Theoretical approaches, empirical evidence, and future directions.
Journal of Management, 32, 991–1022.
Lount, R. B., Jr., & Petitt, N. C. (2012). The social context of trust: The role of status.
Organizational Behavior and Human Decision Processes, 117, 15–23.
Magee, J. C., & Galinsky, A. D. (2008). Social hierarchy: The self-reinforcing nature of
power and status. Academy of Management Annals, 2, 351–398.
Maner, J. K., Gailliot, J. K., Butz, D., & Peruche, B. M. (2007). Power, risk, and the status-
quo: Does power promote riskier or more conservative decision-making? Personality
and Social Psychology Bulletin, 33, 451–462.
Maner, J. K., & Mead, N. (2010). The essential tension between leadership and power:
When leaders sacrifice group goals for the sake of self-interest. Journal of Personality
and Social Psychology, 99, 482–497.
Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of organi-
zational trust. Academy of Management Review, 20, 709–734.
Mayer, D. M., Kuenzi, M., Greenbaum, R., Bardes, M., & Salvador, R. (2009). How low
does ethical leadership flow? Test of a trickle-down model. Organizational Behavior
and Human Decision Processes, 108, 1–13.
Mills, C. W. (1956). The power elite. Oxford, England: Oxford University Press.
Mooijman, M., & Graham, J. (2018). Unjust punishment in organizations. Research in
Organizational Behavior, 38, 95–106.
Mooijman, M., Van Dijk, W. W., Ellemers, N., & Van Dijk, E. (2015). Why leaders punish:
A power perspective. Journal of Personality and Social Psychology, 109, 75–89.
Mooijman, M., Van Dijk, W. W., Van Dijk, E., & Ellemers, N. (2017). On sanction-goal
justifications: How and why deterrence justifications undermine rule compliance.
Journal of Personality and Social Psychology, 112, 577–588.
Mulder, L. B., Van Dijk, E., De Cremer, D., & Wilke, H. A. M. (2006). Undermining trust
and cooperation: The paradox of sanctioning systems in social dilemmas. Journal of
Experimental Social Psychology, 42, 147–162.
Pillutla, M., Malhotra, D., & Murnighan, J. K. (2003). Attributions of trust and the cal-
culus of reciprocity. Journal of Experimental Social Psychology, 39, 448–455.
Rink, F., & Ellemers, N. (2015). The pernicious effects of unstable work group member-
ship: How work group changes undermine unique task contributions and newcomer
acceptance. Group Processes & Intergroup Relations, 18, 6–23.
Sassenberg, K., Ellemers, N., Scheepers, D., & Scholl, A. (2014). “Power corrupts” re-
visited: The role of construal of power as opportunity or responsibility. In J.-. W. van
Prooijen, & P. A. M. van Lange (Eds.). Power, politics, and paranoia: Why people are
suspicious of their leaders (pp. 73–87). Cambridge, UK: Cambridge University Press.
Scheepers, D., Röell, C., & Ellemers, N. (2015). Unstable power threatens the powerful
and challenges the powerless: Evidence from cardiovascular markers of motivation.
Frontiers in Psychology, 6, 720.
Schilke, O., Reimann, M., & Cook, K. C. (2015). Power decreases trust in social exchange.
Proceedings of the National Academy of Science, 112, 12950–12955.
Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2011). False-positive psychology:
Undisclosed flexibility in data collection and analysis allows presenting anything as
significant. Psychological Science, 22, 1359–1366.
Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2013). Life after p-hacking. Paper presented
at the Fourteenth Annual Meeting of the Society for Personality and Social Psychology,
New Orleans, LA.
Stouten, J., De Cremer, D., & Van Dijk, E. (2005). I'm doing the best I can (for myself):
Leadership and variance of harvesting in resource dilemmas. Group Dynamics: Theory,
Research, and Practice, 9, 205–211.
Tost, L. P. (2015). When, why, and how do powerholders “feel the power”? Examining the
links between structural and psychological power and reviving the connection be-
tween power and responsibility. Research in Organizational Behavior, 35, 29–56.
Tyler, T. R., Jackson, J., & Mentovich, A. (2015). On the consequences of being an object
of suspicion: Potential pitfalls of proactive policing. Journal of Empirical Legal Studies,
12, 602–636.
Van Kleef, G. A., Oveis, C., Van der Löwe, L., Luogokan, A., Goetz, J., & Keltner, D.
(2008). Power, distress and compassion: Turning a blind eye to the suffering of
others. Psychological Science, 19, 1315–1322.
Wojciszke, B., & Kujalowicz-Struzynska, A. (2007). Power influences self-esteem. Social
Cognition, 25, 472–494.
Yamagishi, T., & Yamagishi, M. (1994). Trust and commitment in the United States and
Japan. Motivation and Emotion, 18, 129–166.
Zand, D. (1997). The leadership triad: Knowledge, trust, and power. New York: Oxford
University Press.